Synthetic Data Training Programme Now Fully Available Online
We are delighted to announce that the complete recording of our joint educational programme on artificial intelligence and its applications in haematology, organised in collaboration with ERN-EuroBloodNet throughout May 2026, is now fully available online. All four webinar sessions are open to researchers, clinicians, and data scientists worldwide, providing free access to cutting-edge resources designed to advance collaborative research in rare blood disorders.
Session 1: SYNTHEMA – Infrastructure for SDG in ERN-EuroBloodNet
The opening session introduces the core clinical challenges of data scarcity and fragmentation in rare diseases. Our experts outline the strategic vision of the SYNTHEMA project and demonstrate how federated learning and synthetic data generation can safely bypass privacy barriers. You will also see a live demonstration of the platform interface, showing how researchers can train generative models locally without moving sensitive patient records from their clinical sites.
Speakers: Maria del Mar Mañú Pereira and Federico Alvarez (with a platform demonstration by Santiago Barrio)
Session 2: Regulatory on Medical Devices, AI Act, Privacy-Preserving and Ethical Considerations
This session tackles the complex legal and ethical frameworks governing digital health innovations. The presentations explain how the European Union AI Act classifies diagnostic software under its high-risk tier, aligning compliance with the Medical Device Regulation. The speakers also detail how to measure privacy leakage under the GDPR and demonstrate advanced cryptographic solutions, including differential privacy and peer-to-peer secure multi-party computation.
Speakers: Nathan Lea, Rudolf Mayer, Sofia Tsekeridou
Session 3: Data Standardisation, Synthetic Data Generation and Interoperability
Interoperability is the foundation of any multi-centre research initiative. In this technical webinar, speakers from several European projects discuss data standardisation, explaining how clinical records are mapped to the OMOP common data model. The session covers both tabular and three-dimensional medical image synthesis, illustrating the latest developments in generative adversarial networks, variational autoencoders, and diffusion models across distributed healthcare networks.
Speakers: Sara Reidel, Giulio Spinozzi, Sofia Tsekeridou, Imanol Isasa, Rafael Redondo, Inês Sousa, Jan Ramon, Max Salmi
Session 4: Opportunities of Synthetic Data Generation in Rare Haematological Diseases
The final session focuses on the practical application of these tools to specific clinical use cases. Clinical experts showcase how synthetic data is used to model kidney injury and generate synthetic MRIs to detect silent brain infarcts in Sickle Cell Disease. The second half of the webinar focuses on Acute Myeloid Leukaemia, showing how synthetic cohorts can preserve complex genetic profiles, replicate survival curves, and synthesise histopathological bone marrow scans.
Speakers: Maria del Mar Mañú Pereira, Raffaella Colombatti and Eleonora Iascone
Help Us Spread the Word
Please share these resources with colleagues, researchers or clinical partners who are active in the fields of artificial intelligence, patient privacy and rare disease research. Doing so helps build a stronger, more connected community dedicated to improving outcomes for patients with rare haematological diseases.